
This folder includes files relevant to dynamic IRT estimation using the Martin-Quinn judicial data (Figures 8 and 9),
and has several files:

First, we need to have the relevant estimates from MQ via MCMC.  These are:
	mqData2013.Rda: This is the raw data used for MQ estimates
	estimateForKosuke.R: This is replication code provided by Quinn to conduct MCMC estimation.  THIS CODE TOOK 5.5 DAYS TO RUN ON TUKEY!!!!
	MQout.Rdata: The raw output file from MQout
	justices.csv: The MQ scores reported on MQ's website, used to confirm we get the same estimates

Next, we have code to run the MQ data using the variational model.  These files are:
	mq.R: Code that processes mqData2013.rda into a form suitable for input for variational estimation, then estimates the model
	mq_in.rda: The postprocessed data from mqData2013.Rda, an intermediate output of mq.txt
	mq_out.rda: The output from the dynamic variational model on original MQ data, using input from mq_in.rda
	mq_bootstrap.R: Code for parametric bootstrap of output from mq_out.rda
	bootstrap.rda: Results of bootstrap from mq_bootstrap.R

Finally, we have the code that aggregates all the estimates
	comparisons.txt: This does all the analysis using the files generated above, and makes graphics
	/pdf folder: All the graphics for Figures 8 and 9.  Figure 8 is "corr_by_term.pdf"